Stock-performance-basic
John Yuill
2023-01-14
Get price data
Getting daily price data for any stock is as easy as using quantmod package and getSymbols function.
- fetches data as xts time series object
- note src and updated in str output
- auto.assign:
- TRUE: creates xts object in environment with name of symbol
- FALSE: prints results OR can be used to save results to object of any chosen name
getSymbols("CSIQ", auto.assign=TRUE) ## auto.assign=TRUE saves object## [1] "CSIQ"csiq_prices <- getSymbols("MSFT", auto.assign=FALSE) ##auto.assign=FALSE to save to specified object
## xts object
str(CSIQ)## An 'xts' object on 2007-01-03/2023-01-13 containing:
## Data: num [1:4037, 1:6] 11 10.8 10.5 10.1 10 ...
## - attr(*, "dimnames")=List of 2
## ..$ : NULL
## ..$ : chr [1:6] "CSIQ.Open" "CSIQ.High" "CSIQ.Low" "CSIQ.Close" ...
## Indexed by objects of class: [Date] TZ: UTC
## xts Attributes:
## List of 2
## $ src : chr "yahoo"
## $ updated: POSIXct[1:1], format: "2023-01-14 17:57:45"Entire history of the daily stock price (at least going back to 2007) is now available in an xts object named after the stock symbol.
- default source is yahoo finance (src=‘yahoo’)
- default start date is ‘2007-01-01’ (from=‘2007-01-01’)
Provides daily Open, High, Low, Close, Volume, Adjusted (aka OHLCVA):
## show first rows
head(CSIQ)## CSIQ.Open CSIQ.High CSIQ.Low CSIQ.Close CSIQ.Volume CSIQ.Adjusted
## 2007-01-03 10.98 11.00 10.65 10.82 176300 10.82
## 2007-01-04 10.82 10.82 10.39 10.41 164100 10.41
## 2007-01-05 10.50 10.70 10.12 10.15 68600 10.15
## 2007-01-08 10.10 10.34 9.95 9.98 195500 9.98
## 2007-01-09 10.00 10.05 9.77 9.77 85900 9.77
## 2007-01-10 9.69 9.97 9.61 9.68 77800 9.68## show last rows
tail(CSIQ)## CSIQ.Open CSIQ.High CSIQ.Low CSIQ.Close CSIQ.Volume CSIQ.Adjusted
## 2023-01-06 32.26 36.21 32.16 36.15 2386400 36.15
## 2023-01-09 36.60 39.42 36.60 38.57 2458700 38.57
## 2023-01-10 38.57 42.18 38.57 42.05 2145300 42.05
## 2023-01-11 41.44 43.77 41.29 42.87 2642100 42.87
## 2023-01-12 43.05 43.38 41.49 43.18 1716300 43.18
## 2023-01-13 42.69 44.07 42.36 43.00 1190700 43.00“Adjusted” = prices adjusted for stock splits and dividends (possibly other developments? major new issues? buy-backs?)
Incl. TSX
- Includes stocks on TSX
- Optionally import into specified object.
- Specify src from variety of available data sources
## specify object with auto.assign=FALSE
ogi <- getSymbols(Symbols='OGI.TO', auto.assign=FALSE, src='yahoo')Filter for time period
- When importing: specific period (start date and end date)
getSymbols("CSIQ", from="2018-01-01", to=Sys.Date(), auto.assign = TRUE)## [1] "CSIQ"CSIQ_2018 <- CSIQ
head(CSIQ_2018)## CSIQ.Open CSIQ.High CSIQ.Low CSIQ.Close CSIQ.Volume CSIQ.Adjusted
## 2018-01-02 16.98 17.16 16.88 17.14 736100 17.14
## 2018-01-03 17.18 17.20 16.83 17.07 987600 17.07
## 2018-01-04 17.13 17.16 16.78 17.06 799100 17.06
## 2018-01-05 17.00 17.19 16.90 17.01 768900 17.01
## 2018-01-08 17.03 17.20 16.85 17.10 727100 17.10
## 2018-01-09 17.08 17.09 16.72 16.95 725500 16.95## reset full CSIQ for later
getSymbols("CSIQ", auto.assign = TRUE)## [1] "CSIQ"- filter after import
CSIQ_2020 <- CSIQ['2020-01-01/'] ## from specified date fwd
CSIQ_90d <- tail(CSIQ, n=90) ## last n
## selected period
CSIQ_Aug <- CSIQ['2020-08/2020-08-31'] ## Aug 1-31- filter with first and last functions in xts
## data in last 3 months (calendar mths, data so far)
CSIQ_3_mth <- last(CSIQ, "3 months" )
## combine first & last to get: first week from last (mont recent) year
## - includes the days of week 1 in the year
CSIQ_first_week <- first(last(CSIQ, "1 year"), "1 week")More Time Series Manipulation examples with xts below in separate section.
Get multiple stocks
Each imported as separate object
getSymbols(c("NEP","SPWR"), auto.assign = TRUE)## [1] "NEP" "SPWR"## OR
stocklist <- c("NEP","SPWR","FSLR")
getSymbols(stocklist, aauto.assign = TRUE)## [1] "NEP" "SPWR" "FSLR"Merge multiple stocks
- use merge function to join cols
- max 2 data sets at a time, unless use do.call
## TWO AT A TIME
## first two
renewables <- merge(NEP, SPWR, join='outer', fill=NA)
head(renewables)## NEP.Open NEP.High NEP.Low NEP.Close NEP.Volume NEP.Adjusted
## 2007-01-03 NA NA NA NA NA NA
## 2007-01-04 NA NA NA NA NA NA
## 2007-01-05 NA NA NA NA NA NA
## 2007-01-08 NA NA NA NA NA NA
## 2007-01-09 NA NA NA NA NA NA
## 2007-01-10 NA NA NA NA NA NA
## SPWR.Open SPWR.High SPWR.Low SPWR.Close SPWR.Volume SPWR.Adjusted
## 2007-01-03 24.34185 25.14735 23.87034 24.13883 1364069 24.13883
## 2007-01-04 24.16503 24.16503 23.18271 23.85069 1446833 23.85069
## 2007-01-05 23.81794 23.81794 23.22855 23.56254 543765 23.56254
## 2007-01-08 23.76555 25.13425 23.68697 24.73477 872375 24.73477
## 2007-01-09 24.65619 25.49443 24.49902 25.33071 1042177 25.33071
## 2007-01-10 25.27832 27.35429 25.23903 27.01375 2544746 27.01375## add another one
renewables <- merge(renewables, FSLR, join='outer', fill=NA)
head(renewables)## NEP.Open NEP.High NEP.Low NEP.Close NEP.Volume NEP.Adjusted
## 2007-01-03 NA NA NA NA NA NA
## 2007-01-04 NA NA NA NA NA NA
## 2007-01-05 NA NA NA NA NA NA
## 2007-01-08 NA NA NA NA NA NA
## 2007-01-09 NA NA NA NA NA NA
## 2007-01-10 NA NA NA NA NA NA
## SPWR.Open SPWR.High SPWR.Low SPWR.Close SPWR.Volume SPWR.Adjusted
## 2007-01-03 24.34185 25.14735 23.87034 24.13883 1364069 24.13883
## 2007-01-04 24.16503 24.16503 23.18271 23.85069 1446833 23.85069
## 2007-01-05 23.81794 23.81794 23.22855 23.56254 543765 23.56254
## 2007-01-08 23.76555 25.13425 23.68697 24.73477 872375 24.73477
## 2007-01-09 24.65619 25.49443 24.49902 25.33071 1042177 25.33071
## 2007-01-10 25.27832 27.35429 25.23903 27.01375 2544746 27.01375
## FSLR.Open FSLR.High FSLR.Low FSLR.Close FSLR.Volume FSLR.Adjusted
## 2007-01-03 30.17 30.29 28.20 28.50 487000 28.50
## 2007-01-04 28.11 28.84 28.00 28.75 325500 28.75
## 2007-01-05 28.65 28.65 27.73 28.15 524300 28.15
## 2007-01-08 28.26 29.00 28.26 28.90 317000 28.90
## 2007-01-09 29.00 29.77 28.95 29.27 232800 29.27
## 2007-01-10 29.22 29.67 29.09 29.16 267700 29.16## MULTIPLE (2+) AT A TIME
## set list of symbols
stockrenew <- c("BEP","TERP","VSLR")
## set up a new environment to hold the imported data
data_env <- new.env()
## get symbols
getSymbols(Symbols=stockrenew, env=data_env, auto.assign=TRUE)## [1] "BEP"## create list of objects within the environment
data_list <- lapply(data_env, Cl) ## apply function: Cl=close; OHLCV (not sure how to get all)
renewables2 <- do.call(merge, data_list)
head(renewables2)## BEP.Close
## 2007-01-03 8.686384
## 2007-01-04 8.643752
## 2007-01-05 8.643752
## 2007-01-08 8.515854
## 2007-01-09 8.553157
## 2007-01-10 8.553157Multiple - most efficient
Most efficient code for getting multiple stocks at once.
- use cbind to combine cols in separaet object as you go
## specify stock symbols
stockrenew3 <- c("ENPH","SEDG","RUN","JKS")
## set empty object
renewables3 <- NULL
## loop through each stock, binding cols as you go
for(stocks in stockrenew3){
renewables3 <- cbind(renewables3,
getSymbols(Symbols=stocks,
from='2018-01-01', periodicity='daily',
auto.assign = FALSE))
}
head(renewables3)## ENPH.Open ENPH.High ENPH.Low ENPH.Close ENPH.Volume ENPH.Adjusted
## 2018-01-02 2.39 2.66 2.38 2.60 852700 2.60
## 2018-01-03 2.62 2.64 2.55 2.58 434100 2.58
## 2018-01-04 2.62 2.62 2.43 2.50 547500 2.50
## 2018-01-05 2.52 2.53 2.42 2.47 512100 2.47
## 2018-01-08 2.46 2.49 2.34 2.41 953000 2.41
## 2018-01-09 2.40 2.45 2.35 2.38 948000 2.38
## SEDG.Open SEDG.High SEDG.Low SEDG.Close SEDG.Volume SEDG.Adjusted
## 2018-01-02 37.65 38.200 37.35 38.05 407300 38.05
## 2018-01-03 38.25 39.150 37.85 39.05 908400 39.05
## 2018-01-04 39.55 39.650 36.45 37.35 1013600 37.35
## 2018-01-05 37.35 38.375 37.15 37.90 553100 37.90
## 2018-01-08 38.00 38.523 37.75 38.35 487500 38.35
## 2018-01-09 38.35 38.500 36.65 37.20 666700 37.20
## RUN.Open RUN.High RUN.Low RUN.Close RUN.Volume RUN.Adjusted JKS.Open
## 2018-01-02 5.88 5.98 5.88 5.92 672400 5.92 24.12
## 2018-01-03 5.92 6.14 5.92 6.05 933100 6.05 25.21
## 2018-01-04 6.05 6.09 5.91 6.06 798500 6.06 24.21
## 2018-01-05 6.07 6.09 5.94 6.01 1062000 6.01 24.30
## 2018-01-08 6.00 6.10 5.92 6.05 466100 6.05 24.25
## 2018-01-09 6.03 6.08 5.96 6.01 564000 6.01 23.70
## JKS.High JKS.Low JKS.Close JKS.Volume JKS.Adjusted
## 2018-01-02 25.44 24.12 25.06 396600 25.06
## 2018-01-03 25.21 24.08 24.33 534000 24.33
## 2018-01-04 24.42 24.00 24.19 422800 24.19
## 2018-01-05 24.96 24.07 24.23 328500 24.23
## 2018-01-08 24.50 23.96 24.13 358300 24.13
## 2018-01-09 24.54 23.70 24.00 798600 24.00- CLOSE only: specify [,4] at end of getSymbols
## get CLOSE only: same as above but SPECIFY [,4] at end
renewables3c <- NULL
for(stocks in stockrenew3){
renewables3c <- cbind(renewables3c,
getSymbols(Symbols=stocks,
from='2018-01-01', periodicity='daily',
auto.assign = FALSE)[,4])
}
head(renewables3c)## ENPH.Close SEDG.Close RUN.Close JKS.Close
## 2018-01-02 2.60 38.05 5.92 25.06
## 2018-01-03 2.58 39.05 6.05 24.33
## 2018-01-04 2.50 37.35 6.06 24.19
## 2018-01-05 2.47 37.90 6.01 24.23
## 2018-01-08 2.41 38.35 6.05 24.13
## 2018-01-09 2.38 37.20 6.01 24.00Data frame conversion
## convert xts to data frame - will work with dygraphs
csiq_df <- data.frame(csiq_prices)
## row names will be dates - move to date col
csiq_df$date <- ymd(rownames(csiq_df))
## [opt] set rownames to number as in typical data frame (won't work with dygraphs)
rownames(csiq_df) <- 1:nrow(csiq_df)
## [opt] move date col to first left
csiq_df <- csiq_df %>% select(date,1:(ncol(csiq_df)-1))Visualize
quantmod has its own ‘chartSeries’ / ‘chart_Series’, Dygraphs pkg enables interactive charts designed to work with quantmod right out of the box.
Several options:
- chartSeries: basic charting from Quantmod pkg
- works with xts
- understands .Open, .High, .Low, .Close
- basic version has vol
- easy to add technical layers
- chart_Series: experimental version of basic charting from Quantmod pkg, seems like similar features, different default formatting
- dygraphs: more interactivity, works with quantmod,
xts
- works with data frame as long as rownames are dates
- doesn’t have the technical analysis features (afaik); limited tweaking of format - mostly automatic
- ggplot: if data converted from xts to data frame (example above)
chartSeries: basic line charts (non-interactive)
Candlestick
- chartSeries, chart_Series: single stock, OHLC candlestick by default
## basic chart_Series from quantmod
## - two versions chartSeries (orig) and chart_Series (new experimental)
chartSeries(CSIQ_90d, name="CSIQ - OHLC & vol")
chart_Series(CSIQ_90d, name="CSIQ - OHLC") 
## same:
#chart_Series(OHLC(CSIQ_90d))- Open is either bottom of green bar OR top of red bar
- High is highest point (either line or top of bar)
- Low is lowest point (either line or bottom of bar)
- Close is either top of green bar OR bottom of red bar
Technical features
- SMA: simple moving average
- MACD: moving average convergence divergence signals
- BBands: Bollinger bands
- other features are available in quantmod (not shown)
chartSeries(CSIQ_90d, name="CSIQ - OHLC & Technical Indicators", TA=c(addVo(),addBBands(), addMACD()))
- add more technical with chart_Series - layer in one chart at a time
chart_Series(OHLC(CSIQ), subset='2019-12/2020-07')
add_SMA(n=30, col='blue')
add_SMA(n=7, col='green')
add_MACD() 
add_BBands()
addMomentum()
- same idea all-in-one (leaving out momentum - doesn’t work with chart_Series)
## semi-colon separated TA functions
chart_Series(OHLC(CSIQ), subset='2019-12/2020-07', main="Full Tech Indicators",
TA="add_SMA(n=30, col='blue');
add_SMA(n=7, col='green');
add_MACD();
add_BBands()")
Specific cols
- can isolate specific columns
## two variations
chartSeries(Cl(CSIQ_90d), name="CSIQ - Closing Prices")
## new charting
chart_Series(Cl(CSIQ_90d), name="CSIQ - Closing Prices")
Multiple stocks
- multiple stocks, one chart (but how to identify each stock?)
chart_Series(Cl(renewables3c))
chartSeries(renewables, theme=chartTheme("white"))
Dygraph line/candle (interactive)
- dygraph is designed for time series data - xts OR data frame, with first col as date
- works with quantmod but doesn’t incorporate separate vol or technical features
- for OHLC need dyCandlestick additional command
- (Info on inputs for dygraphs)[https://r-graph-gallery.com/316-possible-inputs-for-the-dygraphs-library.html]
Lines / Candlestick
## use dygraphs pkg
dygraph(OHLC(CSIQ_90d))cndl <- dygraph(OHLC(CSIQ_90d))
dyCandlestick(cndl)Single col
- closing price only
# chart_Series(Cl(CSIQ)) #chart_Series seems to sometimes cause probs with pandoc(?)
dygraph(Cl(CSIQ))Annotations!
Easily add annotations
## shorter date range for readability / relevance
CSIQ_covid <- CSIQ_2020['/2021-05-30']
## find date of peak close for annotation
#max(CSIQ_covid$CSIQ.Close)
peak_date <- index(CSIQ_covid[Cl(CSIQ_covid)==max(Cl(CSIQ_covid))])
## multi-step process
## setup chart parameters
#graph <- dygraph(OHLC(CSIQ_covid), main="Cdn Solar")
## compile/add event annotations
#graph <- dyEvent(graph, c("2020-02-21",as.character(peak_date)),
# c("Start of covid crash","Post-covid peak"),
# labelLoc="top")
## print candlestick
#dyCandlestick(graph)
## pipe together chart info and event annotations individually for more control
dygraph(OHLC(CSIQ_covid), main="Cdn Solar with annotation") %>%
dyEvent("2020-02-21","Start of covid crash", labelLoc="top") %>%
dyEvent(as.character(peak_date),"Post-covid peak", labelLoc="bottom") %>%
dyCandlestick()Show only adjusted close, filtered for 2020 - early 2021
CSIQ_2020 <- CSIQ['2020-01-01/2021-01-31']
graph <- dygraph(Ad(CSIQ_2020), main="Cdn Solar")
graph <- dyEvent(graph, "2020-02-21","Start of covid crash",
labelLoc="top")
graph <- dyEvent(graph, "2021-01-21","Peak after covid crash",
labelLoc="bottom")
graph- just under 1 yr from crash to peak
Adjusted prices
Adjusted price is supposed to adjust for stock splits, while other prices are actual on the trading day.
However…Tesla (TSLA) had a 5-for-1 stock split on Aug 3 dropping price from ~$2,000 to ~$400. Prices collected as of Sep 5, 2020 show post-split throughout.
getSymbols("TSLA", auto.assign = TRUE)## [1] "TSLA"## filter for Aug, pre-split
TSLA_presplit <- TSLA['2020-08/2020-08-30']
tail(TSLA_presplit)## TSLA.Open TSLA.High TSLA.Low TSLA.Close TSLA.Volume TSLA.Adjusted
## 2020-08-21 136.3173 139.6993 135.0033 136.6653 322344000 136.6653
## 2020-08-24 141.7520 141.9333 128.5013 134.2800 300954000 134.2800
## 2020-08-25 131.6593 135.1967 131.2000 134.8893 159883500 134.8893
## 2020-08-26 137.3333 144.4000 136.9087 143.5447 213591000 143.5447
## 2020-08-27 145.3640 153.0400 142.8333 149.2500 355395000 149.2500
## 2020-08-28 153.0080 154.5660 145.7680 147.5600 301218000 147.5600#dygraph(OHLC(TSLA))
## set width and height in pixels
dygraph(Ad(TSLA), width=400, height=300)dygraph(TSLA[,c(4,6)], width=400, height=300)Get RETURNS
- Prices are great, but what REALLY counts is RETURNS!
- Quantmod has functions for different returns by periodicity (daily, weekly, monthly, quarterly, annual)
- periodReturn is the underlying function
- can use allReturns to calculate all return periods at once, depending on date range.
- PerformanceAnalytics pkg is the definitive R pkg for returns and portfolio optimization - so this is quick coverage and will dive more into PerformanceAnalytics in another file.
CSIQ_90d_return <- dailyReturn(CSIQ_90d)
head(CSIQ_90d_return)## daily.returns
## 2022-09-07 0.059738434
## 2022-09-08 -0.069134230
## 2022-09-09 -0.016910376
## 2022-09-12 -0.035564807
## 2022-09-13 0.016148421
## 2022-09-14 0.006166935dygraph(CSIQ_90d_return, width=600, height=250)All Returns data
- one command to calculate returns for all periodicities - assuming date range is wide enough.
CSIQ_returns <- allReturns(CSIQ)
head(CSIQ_returns)## daily weekly monthly quarterly yearly
## 2007-01-03 NA NA NA NA NA
## 2007-01-04 -0.037892791 NA NA NA NA
## 2007-01-05 -0.024975985 -0.07559199 NA NA NA
## 2007-01-08 -0.016748768 NA NA NA NA
## 2007-01-09 -0.021042084 NA NA NA NA
## 2007-01-10 -0.009211873 NA NA NA NAAnnual returns
- zoom in on annual returns.
## subset for annual return data and chart
CSIQ_returns[!is.na(CSIQ_returns$yearly),5] %>%
dygraph(main="CSIQ Annual Returns")head(CSIQ_returns[!is.na(CSIQ_returns$yearly),5])## yearly
## 2007-12-31 1.5637523
## 2008-12-31 -0.7705151
## 2009-12-31 3.4613003
## 2010-12-31 -0.5700902
## 2011-12-30 -0.7853107
## 2012-12-31 0.2781955- would be nicer to have bar chart
- how about quick ggplot after conversion to data frame?
csiq_ret_df <- data.frame(CSIQ_returns[!is.na(CSIQ_returns$yearly),5])
csiq_ret_df$date <- ymd(rownames(csiq_ret_df))
ggplot(csiq_ret_df, aes(x=as.factor(date), y=yearly))+geom_col()+
scale_y_continuous(labels=percent_format())+
labs(title="Yearly Returns", x="", y="annual % returns")+
theme(axis.text.x = element_text(angle=45, hjust=1))
## or simplified x-axis
csiq_ret_df$year <- year(csiq_ret_df$date)
csiq_ret_df %>% ggplot(
aes(x=year, y=yearly))+geom_col()+
scale_y_continuous(labels=percent_format())+
labs(title="Yearly Returns", x="", y="annual % returns")
Stats on returns
summary(CSIQ_90d_return)## Index daily.returns
## Min. :2022-09-07 Min. :-0.0691342
## 1st Qu.:2022-10-07 1st Qu.:-0.0250181
## Median :2022-11-08 Median :-0.0063732
## Mean :2022-11-09 Mean : 0.0003253
## 3rd Qu.:2022-12-11 3rd Qu.: 0.0189893
## Max. :2023-01-13 Max. : 0.1378660sum(CSIQ_90d_return)## [1] 0.0292809ggplot(CSIQ_90d_return, aes(x=CSIQ_90d_return))+geom_histogram()+
geom_vline(xintercept = median(CSIQ_90d_return))+
geom_vline(xintercept = quantile(CSIQ_90d_return, 0.2))+
geom_vline(xintercept = quantile(CSIQ_90d_return, 0.8))+
scale_y_continuous(expand=c(0,0))+
scale_x_continuous(labels=percent)+
labs(title="Daily returns over the period",
subtitle='Lines: 20th percentile, Median, 80th percentile',
x='Daily % chg')
Time Series Manipulation
Leverage xts to work with time series objects (probably warrants its own file)
End points by period
- day, week, month, qtr, yr.
- returns the locations of each last item in the specified type of period, aligned on calendar.
- will include the last item available in the last period (last data point).
- returns a vector starting with 0 and ending with the extent (last row) of your data.
- can also use k to skip: on=‘years’, k=4 returns location of every 4th yr in data (plus last)
## get index number of last day of each year
index_yr <- endpoints(CSIQ, on="years")
## apply to time series to extract selected end points
CSIQ_yr_end <- CSIQ[index_yr]
head(CSIQ_yr_end[,c(2,3,5)]) ## selected cols for space## CSIQ.High CSIQ.Low CSIQ.Volume
## 2007-12-31 29.15 27.26 1461900
## 2008-12-31 6.95 6.37 1024900
## 2009-12-31 29.44 28.81 686600
## 2010-12-31 12.41 12.12 510700
## 2011-12-30 2.70 2.53 350400
## 2012-12-31 3.43 3.25 506200index_week_end <- endpoints(CSIQ_Aug, on="weeks")
CSIQ_Aug[index_week_end, 1:4] ## open, high, low, close selected## CSIQ.Open CSIQ.High CSIQ.Low CSIQ.Close
## 2020-08-07 25.00 26.72 24.62 25.32
## 2020-08-14 26.41 26.41 25.51 25.89
## 2020-08-21 25.59 26.65 25.59 25.96
## 2020-08-28 29.25 32.72 28.92 32.52
## 2020-08-31 32.45 34.10 31.78 32.49Summarization
##
period.apply(CSIQ_Aug, INDEX=index_week_end, FUN=mean)## CSIQ.Open CSIQ.High CSIQ.Low CSIQ.Close CSIQ.Volume CSIQ.Adjusted
## 2020-08-07 23.526 24.352 23.142 23.822 1688280 23.822
## 2020-08-14 25.044 25.778 24.274 25.160 1739680 25.160
## 2020-08-21 26.142 26.854 25.756 26.176 1371340 26.176
## 2020-08-28 27.702 29.188 27.424 28.798 2473340 28.798
## 2020-08-31 32.450 34.100 31.780 32.490 2033100 32.490Shortcut summarization
## apply.weekly, apply.monthly, apply.quarterly, apply.yearly to summarize
## FUN=<whatever summarization function for the period)
apply.weekly(CSIQ_Aug, FUN=mean)## CSIQ.Open CSIQ.High CSIQ.Low CSIQ.Close CSIQ.Volume CSIQ.Adjusted
## 2020-08-07 23.526 24.352 23.142 23.822 1688280 23.822
## 2020-08-14 25.044 25.778 24.274 25.160 1739680 25.160
## 2020-08-21 26.142 26.854 25.756 26.176 1371340 26.176
## 2020-08-28 27.702 29.188 27.424 28.798 2473340 28.798
## 2020-08-31 32.450 34.100 31.780 32.490 2033100 32.490Split into periods
## split data into weeks - chunk the data into 'week' components, without summarizing
CSIQ_Aug_weeks <- split(CSIQ_Aug, f='weeks')
## second week: high, low, close selected
CSIQ_Aug_weeks[[2]][,2:4]## CSIQ.High CSIQ.Low CSIQ.Close
## 2020-08-10 25.48 23.94 24.46
## 2020-08-11 24.72 23.32 24.04
## 2020-08-12 25.73 24.00 25.03
## 2020-08-13 26.55 24.60 26.38
## 2020-08-14 26.41 25.51 25.89Aggregate time periods
- Aggregate data by period
- Can take a single variable and calculate OHLC statistics by period
CSIQ_yearly <- to.period(CSIQ[,6], period='years', name='CSIQ_adj', OHLC=TRUE)
CSIQ_yearly## CSIQ_adj.Open CSIQ_adj.High CSIQ_adj.Low CSIQ_adj.Close
## 2007-12-31 10.82 30.61 7.14 28.15
## 2008-12-31 28.67 51.00 3.78 6.46
## 2009-12-31 6.49 30.15 3.00 28.82
## 2010-12-31 30.99 32.71 9.20 12.39
## 2011-12-30 12.49 16.38 2.20 2.66
## 2012-12-31 2.80 4.39 2.03 3.40
## 2013-12-31 3.40 32.00 3.15 29.82
## 2014-12-31 34.19 43.67 21.11 24.19
## 2015-12-31 24.69 39.09 15.72 28.96
## 2016-12-30 27.98 27.98 10.65 12.18
## 2017-12-29 12.67 18.87 11.15 16.86
## 2018-12-31 17.14 17.75 11.87 14.34
## 2019-12-31 14.44 24.99 14.44 22.10
## 2020-12-31 22.52 53.16 13.24 51.24
## 2021-12-31 52.64 64.15 30.18 31.29
## 2022-12-30 31.33 47.01 22.54 30.90
## 2023-01-13 30.30 43.18 30.30 43.00Shortcut aggregation
CSIQ_2020_mth_agg <- to.monthly(CSIQ_2020[,6], name='CSIQ_adj')
CSIQ_2020_mth_agg## CSIQ_adj.Open CSIQ_adj.High CSIQ_adj.Low CSIQ_adj.Close
## Jan 2020 22.52 23.45 20.05 20.05
## Feb 2020 20.62 24.64 20.40 20.40
## Mar 2020 20.61 21.50 13.24 15.91
## Apr 2020 14.74 18.58 14.74 17.84
## May 2020 17.14 19.28 16.81 18.80
## Jun 2020 18.71 20.20 16.95 19.27
## Jul 2020 19.33 24.88 19.33 22.14
## Aug 2020 22.42 32.52 22.42 32.49
## Sep 2020 33.75 35.10 27.68 35.10
## Oct 2020 36.35 44.12 35.90 36.35
## Nov 2020 36.96 44.17 35.70 42.76
## Dec 2020 41.98 53.16 37.53 51.24
## Jan 2021 52.64 64.15 50.52 54.75Merging data with difference frequencies
## get data -> DFF is daily Fed rate, FEDFUNDS is monthly
getSymbols(Symbols=c("FEDFUNDS","DFF"), src="FRED" )## [1] "FEDFUNDS" "DFF"head(DFF) ## daily data ## DFF
## 1954-07-01 1.13
## 1954-07-02 1.25
## 1954-07-03 1.25
## 1954-07-04 1.25
## 1954-07-05 0.88
## 1954-07-06 0.25head(FEDFUNDS) ## mthly data## FEDFUNDS
## 1954-07-01 0.80
## 1954-08-01 1.22
## 1954-09-01 1.07
## 1954-10-01 0.85
## 1954-11-01 0.83
## 1954-12-01 1.28## aggregate DFF to monthly by average
DFF_monthly <- apply.monthly(DFF, FUN=mean) ## FUN=aggregation calc
head(DFF_monthly) ## date as of end of month## DFF
## 1954-07-31 0.7993548
## 1954-08-31 1.2206452
## 1954-09-30 1.0666667
## 1954-10-31 0.8487097
## 1954-11-30 0.8336667
## 1954-12-31 1.2777419## convert to monthly index
index(DFF_monthly) <- as.yearmon(index(DFF_monthly))
## mow can merge with data sets that have first of month
merged_funds <- merge(FEDFUNDS, DFF_monthly)
head(merged_funds)## FEDFUNDS DFF
## 1954-07-01 0.80 0.7993548
## 1954-08-01 1.22 1.2206452
## 1954-09-01 1.07 1.0666667
## 1954-10-01 0.85 0.8487097
## 1954-11-01 0.83 0.8336667
## 1954-12-01 1.28 1.2777419Quandl for Importing
Claims ’the world’s most powerful data lives on Quandl. See also: Quandl API documentation.
- Economic data and similar
- Specify Source/Symbol
- Several ‘types’ available: deflaut is data.frame
- Does not appear that you can access stock prices without account/payment :(
## as of Jan 14, 2023 getting error msg that I need to get API key
#us_gdp <- Quandl(code='FRED/GDP')
#ggplot(us_gdp, aes(x=Date, y=Value))+geom_line()PerformanceAnalytics [next]
PerformanceAnalytics pkg is considered best for analysis of returns and portfolio decisions. Will dive into it in separate file.